| Detailed contour maps/visualized images produced from LiDAR-derived DEMs are powerful tools to locate landslide areas, thereby improving the quality of landslide inventories. However, information of not only locations but also of slope types and activities are wanted in constructing/maintaining infrastructure to mitigate landslide disasters. Since LiDAR measurement can extract the fine surface features associated with slope activities, the DEMs can also be utilized to provide this information.In this study LiDAR-derived DEMs are analyzed to quantify surface features of colluvial and landslide blocks. Two landslide prone areas were selected for analysis, from Kii Peninsula and Fukushima Prefecture in Japan. In the former Mesozoic sedimentary bedrocks are tectonically shared, while Tertiary tuff dominates in the latter. DEM grid sizes used were 1-m and 2-m, respectively. The eigenvalue ratio, a filter to indicate three dimensional surface roughness, and slope angle were used to characterize slope morphometry. Field investigations were also carried out to examine the relations between filter values and actual surface features, block types, and recent slope activities.Results showed that colluvial blocks were characterized by slope angles that clustered about the angle of repose. In the Kii Peninsula site, a high proportion of eigenvalue ratios between 2.75 and 4 indicated the presence of colluvial slope deposits. In both the study sites, recently active slides were steeper than surrounding areas and contained a higher proportion of cells with eigenvalue ratios < 2.5 (scars and bedrock exposure). The Fukushima site also comprised an active mature slide with an undulating surface. Here, there were a high proportion of cells with an eigenvalue ratio of 4-5.In this study the DEM-derived eigenvalue ratio and slope angles could characterize slope types, features and their activities. However, the relationships between filter values and field conditions were not the same between the sites, because of differences in the dominant hillslope processes, and cell-size and resolution of DEMs. Therefore supplementary field investigations are always necessary in the routine application of these filters to characterize landslides and slope features elsewhere. Furthermore, a number of different roughness filters may be required in addition to the eigenvalue ratio to help fully discriminate between surface features in other land settings. |
